Wei Gao, Qinghua Li, Bo Zhao and Guohong Cao The Pennsylvania State University From: MobiHoc 2009 Presentor: Mingyuan Yan Advisor: Dr. Yingshu Li
1 2 3 4 5 Introduction Problem definition Single-data multicast Multiple-data multicast Performance evaluation 6 Related work 2
Introduction 3
Carry-and-forward methods Motivations End-to-end connections are hard to maintain due to low node density and unpredictable node mobility Mobile nodes are used as relays to carry data Main problem: how to determine appropriate relay selection strategy and forwarding criteria Difference between multicast and unicast A relay is chosen to forward data to as many destinations as possible We need to calculate the cumulative probability to forward data to multiple destinations (may require global knowledge of social relations among nodes) Social relations among mobile users are more likely to be long-term characteristics and less volatile than node mobility 4
Social Network Perspective Two concepts in SNA (Social Network Analysis) methods Communities : formed according to people s social relations Centrality: some nodes in a community are the common acquaintances of other nodes and act as communication hubs Properties Contacts vs. mobility Social relations: stable, long-term characteristics, another form of mobility regularity Social-based approaches Majority: unicast Semantic multicast models : SimBet [2], BUBBLE Rap[3] 5
Objective Multicast: improve cost-effectiveness by effective relay selections Minimize the number of used relays Satisfy the required delivery ratio and delay 6
Contributions Analytical models for relay selections Single-Data Multicast Multiple-Data Multicast Unified knapsack formulation for DTN multicast problems 7
Problem Definition 8
Network model Node with buffer constraint each node N k has a buffer size B k Trivial for SDM Necessary for MDM The node buffer may be only enough to carry a part of the data items Which data item to carry? Key difference between SDM and MDM! 9
Single-Data Multicast (SDM) Deliver a data item to a set Problem definition Multiple-Data Multicast (MDM) of destinations Deliver a set of data items to destination sets, respectively Data items has sizes Choose the minimum number of relays Achieve the delivery ratio p within the time constraint T 10
Delivery ratio Average ratio of data items being delivered to destinations For MDM, The average probability that a destination node receives the data item within time T is higher than p Different from the strict definition For each destination node, the probability that it receives the data item within T is higher than p 11
Problem formulation Basic idea: social-based relay selection metrics Assume the contacts of each node pair as a Poisson process Unified knapsack formulation w k : social-based metric values for mobile nodes W: the totally required metric value determined by the required delivery ratio p and delay T 12
Single-data multicast 13
Single-data multicast Localized centrality-based heuristic approach Centrality metric for weighted social network Relay selection: based on nodes centrality values Ensure that all the nodes are contacted by the data source or the selected relays within time T Assumption Destinations are uniformly distributed in the network 14
Centrality metric 15
Single-data multicast 16
Define random variable Relay selection (in contact) Probability that is not contacted by within T Constraint 17
Relay selection (not in contact) Accurate Lower bound The average probability for the relay choice lower bound has the similar 18
Multiple-data multicast 19
Multiple-data multicast A community-based approach is used Each node maintains its destination-awareness about the other nodes in the same community Inter-community data forwarding is done via the gateway nodes 20
Social Forwarding Path Weight: the probability that a data item is forwarded from A to B within time T PDF: 21
Edge-splitting process Social forwarding paths may be overlapping The probability that S sends data to D within time T is not No analytical form! 22
Edge-splitting process Step 1: Move e 0 to the end of paths Commutatively of convolution 23
Step 2: The overlapping edge is split to r edges The contact rate is also split Cumulative probability Edge-splitting process 24
Two stage relay selection 25
Data item selection 26
Relay selection 27
Performance evaluation 28
Performance evaluation Comparison Flooding-based approach: Epidemic routing Mobility-based approach: Prophet Social-based approach: SimBet and Bubble Rap Performance Delivery ratio Actual delay Average cost 29
Performance evaluation 30
Performance of SDM 31
Performance of SDM 32
Performance of MDM 33
1. W. Gao, Q. Li, B. Zhao and G. Cao, Multicasting in Delay Tolerate Networks: A Social Network Perspective, MobiHoc, 2009 2. E. Daly and M. Haahr, Social Network Analysis for routing in disconnected delaytolerant MANETs, MobiHoc, 2007 3. P. Hui, J. CROWCROFT and E. Yoneki, Bubble rap: social-based forwarding in delay tolerant networks, MobiHoc, 2008 34
Wei Gao, Qinghua Li, Bo Zhao and Guohong Cao The Pennsylvania State University From: MobiHoc 2009 Presentor: Mingyuan Yan Advisor: Dr. Yingshu Li